Desmatamento no bioma amazônia: evidências empíricas a partir de imóveis rurais

Abstract

This thesis comprises two articles that investigate deforestation on rural properties in the Amazon biome. In the first chapter, the causal impact of municipalities’ inclusion in the List of Priority Municipalities (LPM) on deforestation associated with rural properties is analyzed, using geospatial data from the Rural Environmental Registry (CAR), combined with deforestation data and socioeconomic variables, covering the period from 2010 to 2022. The empirical strategy is based on the Generalized Synthetic Control method, which allows for handling staggered policy adoption and unobserved heterogeneity. The results indicate that the LPM was effective in reducing deforestation, with effects that consolidate over time and exhibit significant spatial heterogeneity across municipalities that benefited from the policy. In the second chapter, the relationship between rural credit and deforestation is investigated, as well as how this relationship behaves under different policy contexts: in 2008, when Central Bank Resolution No. 3,545 was enacted and the Soy Moratorium was implemented; in 2014, when the CAR was introduced; and in 2019, when environmental enforcement policies were relaxed in Brazil. Using a dataset containing deforestation and socioeconomic variables for the period from 2008 to 2019, a continuous treatment approach based on the Generalized Propensity Score (GPS) and semiparametric estimators is employed. The results reveal a positive, non-linear relationship between credit and deforestation, indicating that higher levels of financing are associated with larger increases in deforestation. Additionally, the quantile analysis reveals heterogeneity across the distribution, with more pronounced effects in the upper quantiles, suggesting that areas with higher deforestation intensity respond more strongly to credit expansion, while no significant effects are observed in contexts of low anthropogenic pressure. Taken together, the results indicate that deforestation arises from the interaction between public policy instruments operating through distinct mechanisms. While command-and-control policies, such as the LPM, help reduce deforestation, economic instruments, such as rural credit, may—if not properly conditioned—amplify incentives for land-use conversion. Thus, this thesis contributes to the literature by shifting the analysis to the level of rural properties and by providing robust empirical evidence on the mechanisms through which different policy instruments shape deforestation dynamics, offering relevant insights for improving the design and coordination of environmental policies in the Amazon context.